Tool

Rate limiter playground: token bucket, fixed window and sliding window side by side

The same "30 per second" limit behaves very differently depending on the algorithm. Watch it happen.

Simulation

Limit (requests per window)

Window (seconds)

Bucket size (burst)

Traffic pattern

Average offered rate (per second)

Duration (seconds)

Each row is one algorithm. Green ticks passed, red ticks were rejected (HTTP 429).

Fixed window: simple, but lets up to twice the limit through around a window boundary.
Sliding window log: precise, at the cost of storing one timestamp per request.
Token bucket: allows short bursts up to the bucket size, then a steady refill rate. Many platform APIs behave like this.